基于改进Hough变换的公路车道线快速检测算法

针对机器视觉导航中道路引导线检测方法复杂、实时性不能满足使用要求的问题提出了基于改进Hough变换的公路车道线快速检测算法:根据车道线与路面颜色的不同判断车道线位置计算出车道线的方向候补点群,根据车辆所处车道的情况对方向候补点群进行聚类,以每一类的中点为基准点使用基于一点的改进Hough变换算法计算出车道线的斜率。实验结果表明,一幅图片的处理时间约85 ms,处理结果与实际情况吻合。算法能够检测出高速公路上所有车道线的斜率,为道路引导线的实时检测提供了一种新思路。

Fast detection of lines on highway based on Improved Hough Transform

In this paper,an approach for fast detection of lines on highway based on the Improved Hough Transform was given.The position of the carriageway was determined by the color difference between the carriageway and the road surface and the substitutive points on the carriageway were calculated,which were classified into groups according to a vehicle position on the road.The middle point of each group was regarded as a datum mark,based on which the carriageway slope was figured out by the Improved Hough Transform method.Compared with the traditional Hough transform,the new approach has the features of higher speed,smaller memory space and higher precision.